Claude Code, start to finish
Twelve parts on the coding agent that lives in your terminal — install it, learn how it thinks, then push it with subagents, hooks, MCP and plugins.
The parts are in order — start with the first one.
The Complete Guide to Claude Code, Start to Finish
The full guide to everything in Claude Code — the six stages, the core tools, and how it all fits together. In simple words, with diagrams.
Read article →What Claude Code Is — and How to Install and Start It
Post 2: what Claude Code actually is, where it runs, and the four steps to install it and open your first session. Plain words, checked against the official docs.
Read article →The 4 Basics of Claude Code — CLAUDE.md, Skills, Context, and Tools
Post 3: the four ideas everything else is built on — CLAUDE.md, skills, context, and tools. In plain words, checked against the official docs.
Read article →Give Claude a Teammate: Subagents
A subagent handles a side task in its own context window and returns only the result — so your main chat stays clean. Here is what it is and how to make one.
Read article →Hooks — Make Claude Code Run Your Rules Every Time
Day 5: hooks are shell commands that run automatically at lifecycle events like PreToolUse and PostToolUse — deterministic rules the model cannot skip. In plain words, checked against the official docs.
Read article →Managing Context and Cost in Claude Code
Day 6: context is your real budget. Claude Code sends your whole conversation with every message, so learn /clear, /compact, /context, caching, and thinking effort to stay fast and cheap. Checked against the official docs.
Read article →MCP — How Claude Code Connects to Your Real Tools
Day 7: MCP, the open standard that connects Claude Code to your real tools — Jira, GitHub, Slack, and your database. In plain words, checked against the official docs.
Read article →MCP Deep Dive — Connecting Claude Code to Your Real Systems
A community-requested deep dive: subagents delegate the work, but MCP is the bridge that connects Claude Code to Jira, Slack, databases, and monitoring. Transports, config as code, scopes, auth, and security, checked against the official docs.
Read article →Plugins, Output Styles, and Scheduling in Claude Code
Bundle your setup as a plugin, change how Claude replies with output styles, and run jobs on a schedule with routines, /loop, and Desktop tasks.
Read article →Scaling Claude Code: From One Session to a Whole Team
The final stage of the roadmap. How I scale Claude Code beyond one chat window — headless mode for CI, git worktrees for parallel work, agent teams, background agents, and the opusplan model setting.
Read article →How I Use Claude Code Every Day: Plan, Act, Verify
Day 11: how I actually use Claude Code day to day — the plan, act, verify loop, permission modes including plan mode, and managing sessions with resume and rewind. Checked against the official docs.
Read article →The Agentic Loop: How It All Ties Together
The finale. Every piece from this roadmap — CLAUDE.md, skills, subagents, hooks, MCP, and models — is one part of a single loop: plan, act, verify, repeat.
Read article →